Gravitational Waves Reveal the Universe before the Big Bang: An Interview with Physicist Gabriele Veneziano
It's not usually put like this, but the discovery of primordial gravitational waves two weeks ago has given us our first direct glimpse of a period before the big bang. [The discovery was later retracted, but I'll keep this post here to show why theorists at the time found it so exciting.] The term "big bang" is sometimes taken to mean the beginning of the universe, and that's the impression you get from diagrams such as the one above, which the BICEP2 team showed during the press conference announcing its discovery.
But cosmologists don't know whether the universe had a beginning. The term "big bang" really refers to the beginning of the universe as we know it—that is, an expanding universe filled with matter that has cooled and coagulated into galaxies. Cosmic inflation, the process the BICEP2 results appear to have vindicated, occurred before the big bang by this definition. The universe during inflation was a deeply alien place, devoid of matter, governed by primeval ur-forces, and thoroughly quantum.
If confirmed by other experiments such as the European Space Agency's Planck observatory, the waves wash away most competing hypotheses for the prehistorical universe, including the so-called pre-big bang and ekpyrotic scenarios. Scientific American published an article about those alternatives in 2004, written by Gabriele Veneziano, a theoretical physicist who is perhaps best known as the father of string theory. The article included a graph (see below) with predictions—now falsified—for the strength of gravitational waves. (I’d added the BICEP2 result to this graph; note that it’s slightly higher than what was predicted for inflation at the time the article was published.)
I caught up with Veneziano in his office at New York University last Friday and, if he was disappointed, he hid it well. Here is an edited version of our conversation.
What's the problem with common usage of the term "big bang"?
People are confusing two things in my opinion. There is something which we may call an effective big bang. We have known for quite a while that the universe had to be hot at some time, because without that hot moment, we would not have been able to have big-bang nucleosynthesis and all sorts of processes that needed this high temperature. That hot moment, according to the modern view, is what must have happened after inflation. So, clearly the effective big bang was not the beginning of time. There was something before that. The "bang" in people's mind is the idea that there had to be a beginning. But we don't know, really, what preceded inflation.
What about the theorems that say there had to be a singularity—a beginning to time?
You can always talk about the hypothetical big bang that should be there if you use general relativity all the way, but the fact you need to apply quantum mechanics to the theory of general relativity invalidates the arguments that there must have been a singularity. I'm not saying you can prove there was no beginning; this is still something which people have not been able to settle. What went on before inflation—if there was a singularity, if there was a beginning of time—is up in the air. There is a certain amount of fine-tuning needed for inflation, and these initial conditions are really still very mysterious.
So, you're saying that we should mark the end of inflation as the big bang?
I would define the big bang as the moment when the temperature reaches it maximum value, right after the end of inflation. Inflation cools down the universe, but then inflation ends and the energy in the inflaton field immediately converts into heat—so-called reheating. And then the expansion makes the temperature decrease it again. So, presumably the temperature went down and up and down again. If I had to identify a special moment, I’d look at when the temperature reached its maximum value and call that the big bang.
There was also something before that moment, and that was inflation. It's very hard to predict the gravitational waves seen by BICEP otherwise. There's also a consistency check. The gravitational-wave spectrum should be slightly "tilted." That means there is a little bit less power at short scales than there is at large scales. One generic prediction of inflation is that the spectrum should be nearly scale-invariant: the same power at every scale. But when one looks more carefully at inflation, if you want it to end at some point, then the spectrum has to have a small tilt. This is what has been observed in the density perturbations seen by Planck. Now, something very similar is expected for the tensor perturbations also. [Tensor perturbations are gravitational waves; the word “tensor” refers to how they are described in Einstein’s general theory of relativity.] In the simplest model of inflation, there is a precise relation between the amplitude of the tensor perturbations and its tilt. This is the next thing to look for.
What new observations do we need to do that?
Look at the waves at different scales. BICEP has a short range of scales—not enough lever arm to see how the spectrum may depend on scale. If Planck will confirm BICEP, it will help since it works at a different scale. According to David Spergel, who gave a very nice colloquium here yesterday, five more experiments are relevant for confirming BICEP, and I imagine that if you combine all six, you'll have a handle on this tilt. This will be pinned down within a year or year and a half. So, if that consistency relation is confirmed, then I think you can say, "Yes, indeed there was a inflationary period before the big bang."
You and others have said that the gravitational waves are evidence of quantum gravity. Can you explain?
Let's forget quantum mechanics for the moment. Say the universe had some initial classical inhomogeneities. Without inflation, these inhomogeneities grow because of gravity, and you wouldn't understand why the universe is so homogeneous today. With inflation, the opposite is true. These inhomogeneities are stretched to scales we cannot see—which is the point of inflation. But then where is the structure we see coming from? That's where quantum mechanics comes in. The generation of perturbations by quantum mechanics happens all along the period of inflation; it's not something you put in at the beginning. You go from something that is completely homogeneous to something which acquires perturbations. Classical perturbations which existed at the initial time were stretched, and not seen today. Quantum perturbations which are produced propagate, and can be seen.
Before BICEP, the perturbations we've seen were the perturbations of the inflaton. The new ones are fluctuations of the gravitational metric; they're really the two polarizations of gravitational waves. It now seems inevitable to apply quantum mechanics to gravity. That’s an interesting development. That forces people to think seriously about how to do quantum gravity. Whether it’s string theory or something else, I don’t know, but we cannot really put it aside and say, “That’s just theory.”
The BICEP results imply a value for the energy scale of inflation of 1016 gigaelectron-volts.
Yes, the size of the tensor perturbations directly gives H [the rate of expansion during inflation] and therefore the energy density. This value turns out to be close to the Grand Unification scale. Inflation happened at a large energy scale, so you may learn physics at very, very short distances from this kind of experiment. Previous data didn't pin down this scale, which goes clearly in the direction of probing some fundamental theory of quantum gravity or other physics going beyond the Standard Model.
Why, if we see multiple polarization states, do we know they're from gravitational phenomenon as opposed to the electromagnetic or some other field?
I don't see a priori why they couldn't be from other fields. One very crucial check that these are really tensor perturbations will be looking at different scales.
Assuming they are gravitational waves, how do we know that those gravitational waves comes from quantum gravity rather than some other process?
Again, a priori there could be other sources of gravitational waves, perhaps cosmic strings. You need to look at the details. Inflation predicts several features—not only the amplitude and the tilt, but also phase coherence. If these are not confirmed, we will need to think about more complicated models. For instance, there is a paper from some of my colleagues in Rome saying that, if you take seriously the observation by BICEP and the non-observation by Planck at two different scales, you rule out a "red" tilt—more power on large scales—to more than three standard deviations. If the spectrum turned out to be "blue"-tilted—more power on short scales—it would not be easy for standard inflation to explain.
What does BICEP mean for your own pre-big bang model?
In that model, we were postulating that you could do without standard inflation. The big bang was instead preceded by a period in which you start from a very simple flat universe and proceed to a more and more curved universe, until at some point string-theory effects forbid you from reaching a singularity, and instead you bounce to our normal phase of expansion. We could reproduce the cosmic microwave background data, but could not get tensor perturbations, so this was our prediction: no tensor perturbations. Our model seems to be disfavored by the data. In a sense, I was happy to see that the simplest model could be proven wrong. The ekpyrotic proponents have also said there should be very small tensor perturbations. These relatively big tensor perturbations exclude quite a lot of models, even conventional inflationary models.
Shifting to the Large Hadron Collider, are you at all worried about the lack of supersymmetric partners?
Before the LHC run, I was saying that I was hoping very much that they’d find supersymmetry, but that if I had to bet money on something, I would bet on finding just the Standard Model Higgs. The indirect tests of the Standard Model were already so good without any extra stuff. I hope I’m wrong and LHC 13-14 will find new physics. I’d much prefer to find something. String theory needs supersymmetry; I think it would also help to understand the smallness of the cosmological constant. But the atmosphere is not enthusiastic—let’s put it that way.
I’m more worried about the future. I’m pretty sure one can find resources for going to much higher energies—people talk about the 100 TeV machine. China seems to be willing to go for it. The question is whether there will be enough motivation in the community for a long-term venture if there is no hint of exciting new physics from the next LHC run.
So, this is independent of political willpower? Even within the physics community?
Even within the community, yes. You have seen how long the LHC took from conception to delivery to finally having the data. It starts to be of the order of the scientific lifetime of a human being. And for the LHC we knew that either we’d find the Higgs or wouldn’t—in any case, it would be interesting. But now we don’t know where is next threshold for new physics.
One of the arguments for TeV supersymmetry was to explain the Higgs mass. If you lose supersymmetry, what do you do with the Higgs?
This is clearly the issue. Some people are now giving up this principle and going more in the direction of anthropic ideas: there are many possible ground states of string theory, for example, and we need one which has suitable parameters.
My feeling is that we have learned from the LHC, LEP [its predecessor], and so on that quantum corrections—the effects that destabilize the Higgs mass—have been seen in the data. For instance, there is the fact that coupling constants are not constant; they depend on distance or scale. But if you look carefully, what we have seen are quantum effects which depend on the large-distance properties of our theory. The Higgs mass and the cosmological constant are very sensitive to the short-distance structure of the theory. My natural reaction is that maybe some of the mystery comes from our ignorance of short distances. String theory is an example of a theory which introduces a fundamental scale and thereby solves this short-distance problem, but I’m also a bit disappointed with string theory, to be honest. It hasn’t provided answers to those crucial questions.
There’s another issue, of course, which is dark matter. There is something which is clearly still missing. That seems to call for some particle beyond the Standard Model. There are still mysteries. Fortunately!